TalkTalk's break-up, in the data: what's really behind the subscriber cliff

TalkTalk's future has been on a knife-edge for years, but the last two weeks have moved the story from "will it find a buyer" to "who ends up holding which piece." Opus Broadband's exclusivity window to buy the ~1.8m-customer consumer arm lapsed on 18 September without a deal, PXC's wholesale sale to Octopus Investments (via Fern Trading) is still in negotiation, and Ofcom has convened emergency contingency reviews, which is the regulator's clearest signal yet that it's preparing for the possibility that one or both processes fail. Sir Charles Dunstone and largest lender Ares Management are now the most likely candidates to take direct control of the remaining consumer book.

Point Topic's GBS subscriber data lets us put real numbers behind that headline, and the picture is more specific, and in one respect slightly more reassuring, than "TalkTalk is disappearing."

The decline was orderly, then it wasn't

For six straight quarters, TalkTalk's retail residential base shrank at a steady clip: down from 3.15m in Q3 2024 to 2.57m in Q1 2026, a loss of roughly 4–5% a quarter. That's a slow bleed consistent with a debt-laden operator losing price-sensitive customers to Altnets and full-fibre rivals, unpleasant, but not seriously alarming on its own.

Then Q2 2026 broke the pattern: 766,000 subscribers gone in a single quarter, a 30% drop that dwarfs anything before it.

Source: Point Topic estimates, company reports, and various news outlets.

The numbers bridge almost exactly to the September figure

That step-change lines up with the deal timeline rather than a sudden wave of churn. Working forward from the Q2 2026 GBS total of 1.80m:

• − 120,000 — accounts transferred to CVC-backed Rise Fibre (3 August)

• − 13,500 — accounts transferred to Fleur Telecom (1 September)

• = ~1.67m implied base

• versus ~1.5–1.6m cited as the current remaining consumer book in the 20 September status update

  
  

The two disposals account for the large majority of the further drop since Q2, leaving only 70,000–170,000 to attribute to ordinary Q3 churn. That's an important distinction: most of the shrinkage since Q2 looks like TalkTalk actively and cleanly offloading accounts as part of a managed wind-down, not customers fleeing in the middle of organisational chaos.

The pain is concentrated in copper, not fibre

Breaking the Q1 to Q2 2026 collapse down by technology tells a sharper story than the headline number:

• FTTx (FTTC/copper-fibre): −516,000 (−46%) — bore the overwhelming brunt

• FTTP (full fibre): −160,000 (−15%)

• DSL (legacy copper): −90,000 (−24%)

  
  

We estimate FTTP held up through all of 2025 and only gave back ground in Q2 2026, while FTTC customers were hit hardest, consistent with a cohort being transferred or exited in bulk from TalkTalk's Openreach copper-fibre base, plausibly overlapping with the exchange stop-sell/PSTN switch-off timetable. We estimate TalkTalk's full-fibre share of what's left has gone from 24% a year ago to 51% today, the more defensible core of whatever Ares and Dunstone end up directly controlling.

Source: Point Topic estimates and company reports when available.

What it would mean if the sales fall through and Ofcom has to step in

1. It would be a first real test of the telecoms Special Administration Regime. Unlike energy, broadband doesn't have a well-worn "supplier of last resort" auction mechanism. The closest tool, the Communications Act's special administration regime, is largely untested at anything near 1.5–1.6m lines. Ofcom convening contingency reviews now looks like pre-positioning for that gap rather than reliance on an established playbook.

2. The risk is concentrated in the legacy-copper cohort, not the network as a whole. The customers most exposed to a disorderly outcome are disproportionately those still on Openreach copper/FTTC, many of whom are already caught in the exchange stop-sell programme. A forced administration landing mid-migration is a far worse moment for them than for the now-majority full-fibre base.

3. The likely default outcome is consolidation, not competition. The Rise Fibre and Fleur Telecom transfers suggest the preferred mechanism (Ofcom's or the administrators') is negotiated bulk transfer to solvent buyers rather than a single cliff-edge failure. That's the less disruptive path for affected customers. Still, it also means TalkTalk's remaining budget-tier base most likely ends up redistributed among a small number of larger or PE-backed challengers rather than staying with an independent value operator, resulting in a net reduction in credible players at the low-price end of the market, however orderly the transfer.

4. PXC is a separate risk axis from the consumer book. PXC is a wholesale layer essentially sitting on top of Openreach rather than its own access network, so a stalled Octopus/Fern Trading deal threatens contractual and financing continuity for the wholesale customers relying on TalkTalk's own network business, which is a different, less visible failure mode than a retail-book transfer, but one Ofcom would need to manage in parallel.

Source: Point Topic GBS subscriber database (published data through 2026-Q2); asset-offloading timeline and September 2026 estimates as reported in the trade and national press.